Complement systems in invertebrates. The ancient ... - CiteSeerX

Immunopharmacology 42 Ž1999. 107–120

Review

Complement systems in invertebrates. The ancient alternative and lectin pathways L. Courtney Smith a

a,)

, Kaoru Azumi b, Masaru Nonaka

c

Department of Biological Sciences and Institute of Biomedical Sciences Graduate Program in Genetics, George Washington UniÕersity, Washington, DC, 20052, USA b Department of Biochemistry, Graduate School of Pharmaceutical Sciences, Hokkaido UniÕersity, Sapporo, 060-0812, Japan c Department of Biological Sciences, Graduate School of Science, UniÕersity of Tokyo, Hongo, Tokyo, 113-0033, Japan Accepted 23 December 1998

Abstract The complement system in higher vertebrates is composed of about thirty proteins that function in three activation cascades and converge in a single terminal pathway. It is believed that these cascades, as they function in the higher vertebrates, evolved from a few ancestral genes through a combination of gene duplications and divergences plus pathway duplication Žperhaps as a result of genome duplication.. Evidence of this evolutionary history is based on sequence analysis of complement components from animals in the vertebrate lineage. There are fewer components and reduced or absent pathways in lower vertebrates compared to mammals. Modern examples of the putatively ancestral complement system have been identified in sea urchins and tunicates, members of the echinoderm phylum and the protochordate subphylum, which are sister groups to the vertebrates. Thus far, this simpler system is composed of homologues of C3, factor B, and mannose binding protein associated serine protease suggesting the presence of simpler alternative and lectin pathways. Additional components are predicted to be present. A complete analysis of this invertebrate defense system, which evolved before the invention of rearranging genes, will provide keys to the primitive beginnings of innate immunity in the deuterostome lineage of animals. q 1999 Elsevier Science B.V. All rights reserved. Keywords: Complement; Alternative; Lectin; Evolution; Echinoderm; Tunicate; Sea urchin; Opsonin

1. Introduction and background The immune response in higher vertebrates is a multilayered complex of interregulated subsystems, including adaptive and innate responses that are mediated by both cellular and humoral systems. The complement system is a major component of immunity in vertebrates and is composed of about thirty distinct humoral and cell surface proteins ŽVolanakis, )

Corresponding author

1998.. The mammalian complement system has three pathways Žclassical, alternative and lectin., that jointly function to amplify an initiating signal through feedback systems of serine protease activities. These pathways converge and activate the central component, C3, which leads to the covalent binding of C3 to the surface of microbes or to immune complexes. One of the central functions of complement activation is to ‘tag’ foreign particles for destruction through augmented phagocytosis of the foreign particle by the phagocyte. In addition, the C3-based tags

0162-3109r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 2 - 3 1 0 9 Ž 9 9 . 0 0 0 0 9 - 0

108

L.C. Smith et al.r Immunopharmacology 42 (1999) 107–120

activate the terminal or lytic pathway that results in the formation of the membrane attack complex. The initiating cascades are regulated by a large number of proteins that are found either in circulation or on cell surfaces. Those in circulation control the rate at which reactive components are formed, and those on cell surfaces function to protect self cells from attack by autologous complement components. The classical pathway is activated by antigen–antibody interactions that bind and activate C1qrC1rrC1s complexes. The alternative pathway is initiated by C3 which undergoes a constant, low-level spontaneous autoactivation reaction enabling it to bind to hydroxyl and amine groups on any protein Žreviewed in Tomlinson, 1993.. The lectin pathway is initiated by mannose-binding lectin ŽMBL. which interacts with mannose sugars on bacterial cell surfaces ŽIkeda et al., 1987. and functions in place of C1q of the classical pathway. MBL-associated serine protease ŽMASP. functions in place of C1s and C1r to activate C2 or C4 of the classical pathway ŽSato et al., 1994. or C3 of the alternate pathway ŽMatsushita and Fujita, 1995; Ogata et al., 1995., although with lower efficiency. The theory that the complement cascades in higher vertebrates evolved from a few primordial genes through gene duplication and subsequent divergence of function ŽBentley, 1988. is based on similarities in protein sequence and function, parallels in complement pathway functions, and clustered organization of some complement genes. Sets of proteins with similar amino acid sequence and function include the thioester protein family ŽC3, C4, C5 and a 2 macroglobulin wa2Mx; Sottrup-Jensen et al., 1985., the C2 and factor B ŽBf. family, and the C1rrC1srMASP-1 and MASP-2 family ŽThiel et al., 1997.. In several cases, duplications and clusters of complement genes have been identified, such as the regulators of complement activation cluster ŽCarroll et al., 1988., the linkage between Bf and C2 genes in human and mouse ŽChaplin et al., 1983; Carroll et al., 1984., the duplication of C4 in humans ŽBelt et al., 1984., and the duplication of C1r and C1s ŽKusumoto et al., 1988; Nguyen et al., 1988.. Finally, parallel functions of the alternative and classical pathways and the similarities between the MBLrMASP complex and the C1 complex are evidence of pathway duplication in the higher vertebrates. The duplication of

ancestral genes and perhaps sets of genes encoding intact pathways has resulted in a complex, multifunctional complement system that is an essential subsystem of the higher vertebrate immune response. Based on a plausible evolutionary history of multiple gene duplications in the complement system of higher vertebrates, it is useful to consider the origins of this system. What ancestral genes were involved in the initial duplication events that lead to the complement system as we know it in mammals? Lachmann Ž1979. proposed that the most primitive complement system would have resembled a simple alternative pathway consisting of a C3-like protein with a thioester site, a factor B-like protein containing short consensus repeats and a serine protease domain, and a complement receptor on phagocytic immune cells. The agnathan complement system was found to fulfill the prediction of a simpler complement system, inferring not only that the alternative pathway was more ancient than the classical pathway, but that it was present in the common ancestor of the vertebrates ŽNonaka et al., 1984.. However, data reviewed here show that an ancestral complement system was present in the common ancestor of the deuterostomes. This indicates that the system may be far more ancient than had been previously believed, and that molecular architecture in the ancestral system may have been somewhat different from the modern alternative pathway as it functions in mammals today.

2. Complement components in the deuterostome invertebrates 2.1. Background The existence of complement components in sea stars and sea urchins was first suggested some time ago. Opsonization of yeast and red blood cells by the mammalian complement component C3 augmented phagocytosis by the circulating cells Žcalled coelomocytes. in echinoderms suggesting the presence of a complement receptor on the coelomocytes ŽKaplan and Bertheussen, 1977; Bertheussen, 1981, 1982; Bertheussen and Seljelid, 1982.. These studies also indicated that a complement-like component in the coelomic fluid could also function in opsonization

L.C. Smith et al.r Immunopharmacology 42 (1999) 107–120

109

and phagocytosis ŽBertheussen, 1983; Leonard et al., 1990.. Although these data were intriguing, these papers were largely ignored due to the lack of molecular data. The molecular basis for the work done by Bertheussen and others and the fulfillment of the predictions by Lachmann has recently been provided in both a sea urchin, Strongylocentrotus purpuratus and a tunicate, Halocynthia roretzi. A characterization of expressed sequence tags ŽESTs. from a cDNA library constructed from sea urchin coelomocytes that had been activated by injection of LPS ŽSmith et al., 1995, 1996. revealed two clones; one homologous to C3 ŽAl-Sharif et al., 1998. and the other to Bf ŽSmith et al., 1998.. Furthermore, homologues of C3 ŽNonaka et al., 1999., Bf ŽJi et al., unpublished. and two MASP proteins ŽJi et al., 1997. have been identified in a tunicate.

appear slightly smaller on gels, consistent in size with that of a mammalian aX chain, indicating that the C3 convertase site may be cleaved ŽAl-Sharif and Smith, unpublished; Nonaka et al., 1999.. In addition, alignments of SpC3 suggest the presence of two putative factor I cleavage sites and activation of sea urchin coelomic fluid results in degradation of the SpC3 a chain into sizes that are consistent with cleavage at one putative factor I cleavage site ŽAlSharif and Smith, unpublished.. Both genes are single copy; SpC3 has a 9 kb message that is present in coelomocytes but not in ovary, testis or gut tissues ŽAl-Sharif et al., 1998., while AsC3 has a 7 kb message that is present in hepatopancreas ŽNonaka et al., 1999..

2.2. C3 homologues in a sea urchin and a tunicate

In mammals, the C3a, C4a and C5a fragments that are released from the N-terminus of the a chains upon activation, are termed anaphylatoxins and possess potent ability to cause smooth muscle contraction, to enhance vascular permeability and to recruit white blood cells ŽEmber and Hugli, 1997.. These activities are dependent on the amino acid sequence at the C terminus of these peptides and substitutions at any of these positions results in a major reduction of activity ŽUnson et al., 1984.. The sequences of the putative SpC3a and AsC3a fragments are atypical in that the C-termini are TSR Žthreoninerseriner arginine. and VSR Žvalinerserinerarginine. respectively, while all vertebrate C3a, C4a and C5a peptides are LXR Žleucineranyrarginine. ŽFig. 2.. In addition, the cysteines in the C3a fragments from the sea urchin and tunicate do not align well to all the conserved cysteine positions in the vertebrate complement fragments ŽFig. 2.. This suggests that the invertebrate peptides may have altered function and conformation compared to the vertebrate peptides, and, as a corollary, the invertebrate receptors for the C3a peptides, if they exist, may also be different.

Expressed sequence tag 064 ŽEST064., isolated from a coelomocyte cDNA library, initially matched to the thioester protein family that includes complement components C3, C4, C5 and a2M. This was the first molecular evidence that an echinoderm had an element within its immune system that was homologous to the complement system, which had been thought to be present only in vertebrates ŽSmith et al., 1996.. In addition, this was the first evidence of homology between immune systems that encompassed the entire lineage of deuterostomes. More recently, a new member of the thioester family has also been identified and characterized from an ascidian ŽNonaka et al., 1999.. Sequence analysis of the deduced proteins from both cDNAs indicated that they are homologues of the vertebrate complement component C3, and have been called SpC3 ŽAl-Sharif et al., 1998. and AsC3 ŽNonaka et al., 1999.. The homologies were based on several conserved regions in both proteins that included Ža. a leader region, indicative of secreted proteins, Žb. a thioester site, Žc. a ba junction and no ag junction, Žd. on reducing protein gels, two chains were present Ž a and b chains of sizes similar to those in mammalian C3 proteins., Že. a putative C3 convertase site, Žf. cysteines in many conserved positions including those involved in forming the interchain disulfide bridge ŽFig. 1A.. Upon activation, a chains of both proteins

2.3. The inÕertebrate C3a fragment

2.4. Phylogenetic analysis of the thioester protein family Comparisons between the sea urchin and tunicate C3 proteins and the vertebrate C3 proteins indicated that the sea urchin and tunicate sequences are equally

110

L.C. Smith et al.r Immunopharmacology 42 (1999) 107–120

Fig. 1. Schematic representation of the structure of invertebrate complement components. ŽA. C3 Structure. The inter-chain disulfide bond, thioester and proteolytic activation site Žarrow. are shown. Neither the a nor b chains show obvious domain structure. ŽB. Domain Structure of Bf. SpBf has the same domain structure as vertebrate Bf and C2 except for two extra SCRs. ŽC. Domain Structure of MASP. The two ascidian MASP proteins have the same domain structure as human MASP-1.

different Žor equally similar. to C3, C4 and C5 proteins and slightly less similar to a 2M sequences ŽAl-Sharif et al., 1998; Nonaka et al., 1999.. Phylogenetic analysis of the thioester protein family showed that SpC3 is the first diverging member of the thioester family of proteins ŽAl-Sharif et al., 1998. and that AsC3 falls within the tree between the echinoderm protein and the vertebrate clade of proteins ŽFig. 3.. This branch arrangement is in agreement with the generally accepted phylogeny of the deuterostomes. 2.5. Bf homologue in a sea urchin The complete sequence of a second clone from the EST study, EST152 ŽSmith et al., 1996., was

found to encode a homologue of vertebrate factor B ŽBf. and was therefore called SpBf ŽSmith et al., 1998.. Like other members of the BfrC2 family, SpBf has a mosaic structure ŽFig. 1B. which includes Ža. five short consensus repeats ŽSCRs. or complement control protein modules, Žb. a von Willebrand factor ŽvWF. domain, Žc. a conserved factor D cleavage site, Žd. a serine protease domain, and Že. Mg 2q binding sites that, in vertebrate Bf proteins, function in interactions with C3b during the formation of the C3 convertase. The gene encoding SpBf is expressed specifically in coelomocytes in the same pattern as that for SpC3. The gene expression data for both sea urchin complement components suggests that the coelomocytes may be a major source of complement production in the sea urchin. This may be because

L.C. Smith et al.r Immunopharmacology 42 (1999) 107–120

111

Fig. 2. C3a peptide alignment. Alignments of the full length proteins were done using ClustalX Žver 1.63b; Thompson et al., 1997., however, only the C3a regions are shown. Slight adjustments in the alignment were done by hand. Accession numbers for the sequences used can be found in the legend to Fig. 3.

these animals do not have the equivalent of a hepatopancreas or a liver, which are the major sources of complement components in tunicates and vertebrates, respectively. A significant difference in the domain structure of SpBf compared to the vertebrate BfrC2 protein family is that the sea urchin protein has five SCRs whereas the vertebrate proteins have three. In mammals, these domains are known to function in interactions between Bf and C3, or between C2 and C4 ŽHoriuchi et al., 1993; Hourcade et al., 1995; Xu and Volanakis, 1997.. To determine whether this difference might have a functional significance, SCR domains from all the BfrC2 proteins including those in SpBf were used as independent sequences in alignments followed by phylogenetic analyses using a variety of outgroups ŽSmith et al., 1998.. Comparisons of the numerous trees that were generated indicated several interesting points. Ž1. The order of the vertebrate and sea urchin SCRs, based on sequence similarities which infers functional similarities, were the same. Ž2. The sea urchin protein organization with five SCRs should be con-

sidered ancestral, and therefore Ž3. the ancestral vertebrate protein may have deleted two SCRs. 2.6. Bf homologue in a tunicate Preliminary analysis indicates that a Bf homologue exists in the tunicate ŽJi et al., unpublished.. Strikingly, the ascidian Bf also has five SCRs in addition to the vWF and serine protease domains. That both the sea urchin and the tunicate homologues have more than three SCRs lends strong evidence that this is the ancestral structure of the Bf protein ŽSmith et al., 1998; Ji et al., unpublished.. The alignment of the amino acids around the catalytic serine residue enables direct comparisons between invertebrate and vertebrate Bf and C2 sequences and has revealed a possible evolution of these proteases ŽFig. 4.. Both Bf and C2 from higher vertebrates lack the aspartic acid that is typically located near the catalytic serine at the bottom of the substrate specificity pocket in vertebrate serine proteases such as trypsin ŽVolanakis and Arlaud, 1998..

112

L.C. Smith et al.r Immunopharmacology 42 (1999) 107–120

Fig. 3. Phylogenetic tree of the thioester protein family. The phylogenetic tree was constructed using the test version 4.0d64 of PAUPU written by David L. Swofford on a Power Macintosh. The a 2M protein family was defined as the outgroup to root the tree. The tree is shown as a phylogram. Reliability of branch lengths were analyzed by 1000 bootstrapping replications and the bootstrap numbers of 50% or greater are shown. The alignment used to generate this tree was done with ClustalX ŽThompson et al., 1997. and was not edited before being imported into PAUPU . Accession numbers for the sequences used in this analysis are: C3 sea urchin, gb